There is a fascinating book called ''Conquer the Casinos'' by Philip Koetsch which I have in my collection. I know a few other members here that have it and also find it interesting. The guy who wrote it showed through his analysis of over 12 million games that you can quit a winner in up to 90% of the rounds of nearly even odds games. His book contained different MM plans for which he tested each one for 100 rounds X 600 = 60,000 games.
His particular favourite was an MM plan called G3M1. He let a winning bet ride three times and bet 2 chips after any 1 chip loss.
Here is a few examples of the G3M1 in action.....
[attachimg=1]
He ran a lot of detailed stats to find out as much information as possible relating to each MM plan. The following is for the G3M1.
[attachimg=2]
The G3M1 was impressive in the sense that it had an average of 2-3 reversals from negative to positive territory. Only 10% of the rounds using G3M1 did not produce a reversal.
So I got to thinking how would this perform with VDW/AP? Naturally 100 games would take longer to play because of the nature of VDW/AP where you are not betting every hand. I will run through some games of 100 bets and compare the stats over the next few weeks. I have enclosed an example below where I stopped when I reached the 'peak average' of 22.8 units. This game was a particularly good run for VDW/AP and only took 18 bets to reach the 'peak average' in 75 hands.
The chart is pretty much easy to follow....
cheers
p.s. The VDW/AP idea I am running through in these tests is just the basic variant. No tweaks like using it on a rolling basis...etc.
What is the definition of a "round" ........seems unclear.
Quote from: 8OR9 on July 03, 2017, 12:55:02 AM
What is the definition of a "round" ........seems unclear.
A round is a hand of cards or spin of the wheel.
The next win took nearly 5 shoes and 78 bets to reach the 3rd reversal. It was a bit of a nightmare run for VDW/AP. A lot of the new games started with a two which would be great for playing on a rolling basis carrying the last decision over......and then it chopped giving out a lot of 'terrible 2's'. On saying that, it only reached -15 at worst.
cheers
Hello Bally!
How do you use or tackle the concept using reversals?
You wrote you reach your third reversal when the average is two, that make no sense, my opinion.
For example:
1) Assume i have a win target of +3 units and push once more to win more and if i don't i stop at +2 units.
2) If i once have a minus i aim to get even once (one reversal) and then again push for my win target, but if i lose and become even and not reach my win target and find my self with minus again i aim to get even - then i have to stop playing or quit session as that was my last reversal.
What i mean is that if you know you get two reversals and hit minus results, from that moment you know you can only get back to even once more to have a chance to try again to hit win target.
And if you lose and are at minus again you know that is your last chance to break even and you can not push to reach your win target again.
So i would like to know how you use the concept of reversals - as i see it - if you don't hit the sweet spot at the beginning or after one reversal you have to end the game and quit.
I mean how would one argue to continue to play for a third reversal?
Many Thanks ...
Cheers
Here is the completed 100 rounds of GM31-2. I will finish the 100 rounds of the first example later today.
The second example (G3M1-2) had 47 wins vs 53 losses and finished up at -1. There were 4 reversals from negative to positive territory.
Hello Sputnik,
I agree with you in a sense because I wouldn't be risking a drawdown just to try and hit a 3rd reversal. Take my first example above as it stands in the opening post. That was plain sailing and hit the average peak figure. I wouldn't have stopped there when the profits are continually rising. I would have kept going until it showed the first sign of a dip. The second completed example was a much tougher game for VDW/AP.
At the moment, all this is just experimental. I want to compare how VDW/AP does against Koetsch's stats using a 'non-random' bet selection. I can pick the bones out of it once I have completed 100 or so games.
cheers
The first game G3M1-1 produced the following.
57 wins vs 43 losses.
End result = +35.
A peak of +41.
A low of -1.
Only one reversal however it was in profit all the way through the shoes apart from the early -1 score.
cheers
Bally i just want to say i am following this topic and i like that you talk about reversals. I try to experiment and include reversal into my own development. Nice work.
Cheers
Thanks Sputnik :thumbsup:
What I should have done is also show the stats for what happens when flat betting.
So here we go.....
[attachimg=1]
And now here are the stats for the G3M1.
[attachimg=2]
You can see some big differences here. For example look at the average % of rounds where net gain exceeds 20 chips with flat betting. It was a paltry 2%. There were no rounds which exceeded 30 chips. Compare that with the G3M1 with figures of 47% and 29% respectively.
cheers
Quote from: Bally6354 on July 03, 2017, 12:56:36 AM
A round is a hand of cards or spin of the wheel.
I was wrong here....sorry! A round is 100 hands or spins. All the tests in the book were conducted over 100 rounds x 600 = 60,000 hands or spins. For anyone interested, there are still some fairly cheap copies floating around on Amazon. Koetsch compiled 20 different MM plans and has compiled all the stats like above and then rates them in order of effectiveness. The G3M1 was his favourite which he claimed to be successful with over a period of several years.
cheers
If i remember correct from the book so is the testing done with 1.7% house edge and is about EC bets. So the expectation would be same with roulette using La Partage Rule (1.32%) I also assume if you play Baccarat and only bet Player you would have a lower house edge then 1.7%.
Cheers
He is using a 1.4% house take which he says is consistent with Baccarat and with the pass-line and come bets in Craps.
cheers
Bally,
What is a reversal and how does he define it?
Mike i will try to explain and Bally might comment.
When you play you can get down with negative results with out any positiv gain from the beginning placing your first wager and never get back to even or zero - then you have a round with no reversal.
All samples that go into negative results and back to even once is one reversal. So each time you go from profit or zero value into negative results and back to zero or break even you have one reversal.
Illustration:
+1 +2 +1 +0 -1 -2 -3 -2 -1 -2 -1 +0 (one reversal) then you might go into profit or hit negative results again and if you then break even again after a negativ result you would have two reversals.
+1 +2 +1 +0 -1 -2 -3 -2 -1 -2 -1 +0 -1 -2 -1 -2- -3 -4 -3 -4 -3 -2 -1 -2 -1 +0 (two reversals).
Cheers
Quote from: Mike on July 03, 2017, 03:13:41 PM
Bally,
What is a reversal and how does he define it?
Hello Mike,
I will quote from the book....
''One of the things I was especially interested to learn was just
how often in a 100 game round will the net-loss cycle typically reverse? And does it always reverse?
Figure 3.3 illustrates just how that pendulum did swing in 600 rounds of 100 random games.
figure 3.3 (This is the flat betting example)
[attachimg=1]
Again, I am only considering a reversal to mean when a net-loss situation reverses to become a net-gain. Considering the theoretical possibilities, the worst case would be absolutely no reversals. But in fact, what we see is that the maximum rounds with no reversals was 17%. This is really pretty good news. 83% of the time at least one reversal will occur. If you quit after that occurrence, you would walk away with the casino's money. Isn't that our real goal?
Still more important is finding that, in a 100 game round, the average number of reversals per round was 3.48! And that 68 of the 100 rounds had 2 or more reversals. What this means is that you can continue to play in a losing round with the statistical knowledge that on the average there will be more than 3 reversals. And that in more than two-thirds of the rounds there will be an average of at least 2 reversals. On the other hand, if one or more reversals have already occurred, continuing to play is statistically risky.''
cheers
Thanks Bally & Sputnik. So do your results show that these stats have any value in actual play?
I think the first thing I need to do is see if I can replicate Koetsch's stats. So I am going to compile 100 rounds of 100 bets and compare the two. It's a bit of a bold claim to say that you can quit a winner in 90% of (100 bet) rounds. Then I suppose it's a case of seeing if the 10% of losing rounds add up to more than the 90% of winning rounds on average. So I thought of giving it a helping hand testing with the VDW/AP which in all my testing so far produces lower variance than a random bet selection. The G3M1 looks like a good fit.....but it's early days.
cheers
Ok final question - what is VDW/AP?
Quote from: Mike on July 03, 2017, 06:56:58 PM
Ok final question - what is VDW/AP?
The following thread pretty much explains it on the first page. Nicksmi kindly uploaded a spreadsheet so anyone can play around with it and see how it all works.
http://betselection.cc/roulette-forum/use-math-to-beat-roulettebaccarat/
cheers
Here is G3M1-3.
End result = -12.
A peak of +10.
A low of -13.
3 reversals.
cheers
Quote from: Bally6354 on July 03, 2017, 07:06:06 PM
The following thread pretty much explains it on the first page. Nicksmi kindly uploaded a spreadsheet so anyone can play around with it and see how it all works.
http://betselection.cc/roulette-forum/use-math-to-beat-roulettebaccarat/
cheers
Thanks. You say that this bet selection has produced a lower variance than a random bet selection. Can you quantify this ? One way of doing it is to count the gap lengths between hits for the VDW and a random selection. Then take the standard deviation of the gaps in both cases. If you're correct, the Stddev of the VDW selection should be lower.
For those struggling with Standard Deviation here is an excellent, step-by-step explanation:
http://www.mathsisfun.com/data/standard-deviation-formulas.html (http://www.mathsisfun.com/data/standard-deviation-formulas.html)
I have found it most useful in coding SD calculations in several of my programs.
Thanks guys for the tips regarding calculating the SD with regards to variance.
Regarding VDW/AP.....there are a couple of interesting threads going on at the moment on the German Forum.
https://www.roulette-forum.de/forum/115-einfache-chancen/
just use google translator to get the gist of it. Loads of charts with info and different ways to attack.
G3M1-4 results....
End result = -1.
A peak of +2.
A low of -21.
2 reversals.
I will do a more detailed analysis compared to Koetsh's stats every 10 games.
cheers
The stats are pretty meaningless after only 4 games, but here they are anyhow.
[attachimg=1]
Regarding the variance, I am going to keep my eye on the average of the peak loss.
Koetsch's stat was 24.9 where I am at 12.5 at the moment.
cheers
G3M1-5 results.....
End result = +46
A peak of +46
A low of -4.
1 reversal.
cheers
Quote from: Bally6354 on July 04, 2017, 06:50:00 PM
G3M1-5 results.....
End result = +46
A peak of +46
A low of -4.
1 reversal.
cheers
Correct me if i am wrong, but i see two reversals. You start from no bet and fall into negative results and then get to break even and then once more into negative result and then again back to positiv results. That count as two reversals as you fall into the negative area twice and get back to even the first time and positiv result the second time. It does not matter if the second reversal only is -1 and then back to even and positiv results. Each negative to even is one reversal.
Cheers
Hello Sputnik,
The way I have always interpreted it is that if you are in plus, go negative and then go plus again = 1 reversal.
So + - + = 1 reversal. So strictly speaking according to my rules, I could say there was no reversal in G3M1-5, however, it started out losing to begin with and then went positive. I do see how you are interpreting it based on what you wrote.
Koetsch writes....''I am only considering a reversal to mean when a net-loss situation reverses to become a net-gain.''
I never was good at following instructions! :))
cheers
Bally are you saying i am wrong or right. Assume you fall into negative results from beginning and then break even, then based upon my understanding that is one reversal.
Does situation where he state there has not been one reversal present at all is does rounds where the result is negative from the beginning and never climb back to even during the 100 trail sample.
Based upon that i see any minus to even as one reversal.
Bally i read what you wrote above, but don't fully understand if you agree or not.
Cheers :thumbsup:
If I take literally what Koetsch wrote, then I interpret it as the following.....
-1, -2, -1, 0, +1 is a reversal because he writes 'a reversal means where a net-loss becomes a net-gain.'
So if it goes....-1, -2, -1, 0.....I don't think that's a reversal because that's not a net-gain. You started with zero profit and still have zero profit.
My original interpretation of needing to be in profit, then losing and then going back into profit was wrong (I think)
So you were closer than me anyhow! :thumbsup:
I will readjust the reversals according to the rules. Thanks for bringing it up because obviously it's important to go by the author otherwise we are playing our own game.
cheers
Updated chart.....
[attachimg=1]
cheers
I understand the betting strategy of betting 2 after a loss. But how would you handle multiple losses in a row, i.e. 4 or 5 in a row? Would you sit out or continuing betting?
I do love the strategy though.
John
How would you handle losing several in a row? Would you continue betting or sit out?
Thanks,
John
(Sorry, I just duplicated the message.)
Results for G3M1-6
End result = +23
A peak of +25
A low of -9
2 reversals
cheers
Hello Bacply,
Suppose I had 5 bets and all were losers.....It would look like this.
Bet 1) x 1 unit.....loss -1
Bet 2) x 2 unit.....loss -2
Bet 3) x 1 unit....loss -1
Bet 4) x 2 unit....loss -2
Bet 5) x 1 unit....loss -1
So that would hit me for 7 units. It happens! I think the longest losing run so far is 5 or 6 hands in these particular tests. (It can go more than that of course)
cheers
Updated chart....
[attachimg=1]
cheers
The G3M1 stands for Geometric x 3 and the M1 part is Martingale x 1.
Koetsch's book covers flat betting
Linear betting....example: 1,2,3... (positive)
geometric betting.....example: 1,2,4,8,16... (positive)
martingale betting....example:1,2.... (negative)
He combines the above together in loads of different ways to come up with his stats.
Bally i just think about something and it will not work, but is a funny tougth, assume we get so many reversals at each game, more then the losing ones without reversals, then why not just wait until one shoe is losing and start betting when they reach zero, as they should make a - net gain - as reversals. Not sure how that would work out, but is a funny thinking process in theory :-)
Now when i think about it, all reversals is about variance getting back to a natural state after a negative state, so we get imbalance and balance with each reversal. So in theory some one could explore reversals that way.
Cheers
can someone please give me a betting example on how to use this live in a casino? use this if you could:
B PP BBBB P B PP B P BB PP BBB PPPPPP B PP B P BB PPP B P
THANKS
This is what it would like...
[attachimg=1]
cheers
thank you, Let me go through this and if confused i will respond here!
No problem!
As far as using the VDW/AP....I always find using this little chart helps!
[attachimg=1]
It's also worth noting that you can't get a 'mutual' bet before the 7th spin using the arithmetic progression.
So for example... looking above at the first appearance of a mutual arithmetic progression on the 9th spin....
3) B
6) B
and
5) P
7) P
So 3,6,9 would be a B and 5,7,9 would be a P. Obviously you can't play them both, so you start a new arithmetic progression.
cheers